Lysis of Gram-Positive and Gram-Negative Bacteria by Antibacterial Monolith Formed in Microﬂuidic Biochips for Sample Preparation

Home , Lysis

Lysis of Gram-positive and Gram-negative Bacteria by Antibacterial Monolith Formed in Microﬂuidic Biochips for Sample Preparation

Mohamed Aly Saad Alya, Mario Gauthierb, and John Yeow∗a

ABSTRACT: Bacterial cell lysis is demonstrated using polymeric microfluidic biochips operating via a hybrid mechanical shearing/contact-killing mechanism. These biochips are fabricated from a cross-linked poly(methyl methacrylate) (X-PMMA) substrate by well-controlled, high-throughput laser micromachining. The unreacted double bonds at the surface of X-PMMA provide covalent bonding for the monolith, thus contributing to the mechanical stability of the biochip and eliminating the need for surface treatment. The lysis efficiency of these biochips was tested for gram-positive (Enterococcus saccharolyticus and Bacillus subtilis) and gram-negative bacteria (Escherichia coli and Pseudomonas fluorescens), and confirmed by off-chip PCR without further purification. The influence of the flow-rate when pumping the bacterial suspension through the monolith, and of the hydrophobic-hydrophilic balance on the cell lysis efficiency was investigated at a cell concentration of 105 CFU/ml. The lysis efficiency of the biochips was better than for off-chip chemical, mechanical, and thermal lysis techniques. The biochip also acts as a filter that isolates cell debris and allows PCR-amplifiable DNA to pass through. As expected, the system performs more efficient lysis for gram-negative than for gram-positive bacteria. The biochip does not require chemical/enzymatic reagents, power consumption, or complicated design and fabrication processes, which makes it an attractive on-chip lysis device that can be used in sample preparation for bio-genetics and point-of-care diagnostics. The biochips were reused for 20 lysis cycles without significant performance degradation or carryover when they were back-flushed between cycles. The biochips efficiently lysed both gram-positive and gram-negative bacteria in about 35 min per lysis and monolith regeneration cycle.

Introduction Electrical cell lysis is reagent free, faster and less expensive than chemical treatment; thus it is widely integrated in micro- Cell lysis refers to a process in which the cells are broken electromechanical system (MEMS) devices. Unfortunately it down by disrupting their membrane and results in the libera- requires a high power consumption, which is an unfavorable tion of their cytoplasmic content, namely the DNA, RNA, and parameter within miniaturized systems. Electrical cell lysis proteins essential for molecular genetic analysis and point- also often demands complicated designs to accommodate the of-care diagnostics. Cell lysis, as a primary step in sample electrodes, which results in difficult fabrication. preparation, has been achieved through fundamentally differ- Thermal cell lysis is accomplished by thermal shock con- ent techniques including electrical, thermal, mechanical and sisting of freezing and thawing cycles, or otherwise by heat- chemical methods. Developments in lab-on-a-chip (LOC) ing a cell suspension, 3 This causes the proteins within the cell techniques have opened up various possibilities to scale down membrane to be denatured, which leads to irreparable damage bio-genetics and molecular diagnostic machines into minia- and release of the intracellular components. As an alternative turized devices with numerous advantages such as higher sen- to chemical or enzymatic lysis, thermal cell lysis is useful to sitivity and accuracy, and a reduction in the reaction vol- avoid contamination in downstream processes. Precise con- umes, which in turn minimizes the cost of the reagents and trol of the temperature to avoid the denaturation of DNA is the amount of waste generated. a significant challenge in thermal cell lysis. 4 The high power Electrical cell lysis can be achieved by exposing the cells requirements for heating also make thermal cell lysis undesir- to a high-intensity pulsed electric field (PEF) 1 that destabi- able for portable systems. lizes and disintegrates the membranes, by producing nano- pores leading to their dielectric breakdown. 2 The lysis effi- In chemical lysis, lysing agents can be used to target the ciency and time both depend on the electric field strength. cells either through continuous flow in microfluidic chan- nels5 or by incubation in micro-chambers 5 or by incuba- 6 a Department of Systems Design Engineering, University of Waterloo, tion in micro-chambers. Buffers such as ammonium chlo- 200 University Ave West, Waterloo, Ontario, Canada N2L 3G1. E-mail: ride, known as Red Blood Cell Buffer, 7 detergents (disrupt- [email protected]; Fax: +1 519 746-4791; Tel: +1 519 888-4567 Ext. ing the cell membranes by solubilizing the membrane proteins 32152 8 b and lipids, thus creating pores) including Triton X-100 and Department of Chemistry, University of Waterloo 9 The final publication is available at Springer via sodium dodecyl sulfate (SDS), chaotropic agents like guani- http://dx.doi.org/10.1007/s00216-014-8028-9 dinium thiocyanate (GTC), 9 ethanol and magnesium chlo-

1 ride (interfering with intermolecular forces in proteins) and lus subtilis (B. subtilis) without requiring chemical/enzymatic enzymes such as lysozyme, 10,11 lysostaphin and proteinase- reagents, power consumption, or complicated design and fab- k11 (digesting the peptidoglycans in the cell walls, ultimately rication processes, making it an attractive on-chip cell lysis compromising their integrity) are the most common chemical technique. 18 We now expand the scope of this work con- agents used. Possible interference of the chemical and enzy- siderably to enhance the cell lysis efficiency, by tuning the matic agents used with the subsequent analysis processes is hydrophobic-hydrophilic balnce of the PPM columns. The the major downside of chemical lysis. optimal flow rate, at which the bacterial cell walls are suffi- Mechanical cell lysis is conceptually the simplest method ciently mechanically sheared through the porous PPM column to achieve lysis, as it uses mechanical forces to disintegrate to disrupt the cell membrane by physical contact with the an- the membrane. There are numerous techniques to disrupt the tibacterial polymeric biocide covering the pore surface, was cell membrane through physical contact between the cells and also determined. The usefulness of this new technique was lysing objects. One method is to force the cells through nar- also further confirmed in terms of reusability of the monolithic row gaps or pores with sharp edges, that are smaller than columns, by demonstrating that the PPM columns suffer from the cells, 12 thereby shearing the cell membrane and ruptur- no significant performance loss when used in successive ly- ing it to release its intracellular content. Burke et al. 13 thus sis cycles. The biochips efficiently lysed both gram-positive formed a narrow porous polymeric monolith (PPM) within and gram-negative bacteria, producing cell lysates contain- a poly(dimethylsiloxane) (PDMS) microfluidic channel and ing DNA that could be amplified by off-chip PCR without used it to lyse white blood cells (B lymphocytes), by shearing the need for purification, which proves that the monoliths do the cell membrane through the small pores of the monolith. not leach PCR inhibitors making them unsuitable for sample Mahalanabis et al. 14 also developed a microfluidic chip to lyse preparation. In this investigation, we also added Pseudomonas bacterial species and extract genomic DNA by mechanical fluorescens (p. fluorescens) and Enterococcus saccharolyti- shearing through a PPM column within a microfluidic chan- cus (E. saccharolyticus), to further expand lysis testing of the nel, but with the assistance of chemical reagents. Blockage antibacterial columns with gram-negative and gram-positive generally remains a drawback in such minute flow-through de- bacteria, respectively. vices. Antimicrobial polymers are materials possessing antimicrobial properties, i.e. the ability to inhibit the growth and eventu- ally kill microorganisms such as bacteria and fungi. Research Experimental is ongoing to engineer these polymers so as to imitate natu- ral host defense peptides (HDPs), used by the immune system in living organisms to kill bacteria. An emerging class Materials of antimicrobial polymers, termed ‘synthetic mimics of antimicrobial peptides’ (SMAMPs), 15 was synthesized to mimic Methyl methacrylate (MMA, 99 %), butyl methacrylate the main features of HDPs: cationic charge and amphiphilic (BuMA, 99 %), ethylene glycol dimethacrylate (EGDMA, character, which facilitate the permeation and disintegration 98 %), and 1,6-hexanediol dimethacrylate (1,6-HDDMA, of bacterial membranes. An antimicrobial surface is a form of ≥90%) were all purchased from Sigma-Aldrich (Oakville, ON antimicrobial polymer killing cells by contact. Tiller et al. 16 Canada), and passed through alumina columns to remove in- introduced in 2001 surfaces that killed bacteria upon contact, hibitors. 1-Dodecanol (98 %, reagent grade), cyclohexanol and termed it ‘contact-killing’. Wan et al. 17 further described (99 % Reagent Plus), 2,2-dimethoxy-2-phenylacetophenone an integrated microchip system allowing cell lysis and PCR (DMPAP, 99 %), methanol (≥99.9 % Chromasolv), fumed within the same reaction chamber. They used gold nanopar- silica (powder, 0.2-0.3 µm average particle size), phosphoric ticles modified with an antibacterial polymer on their surface acid (85%), ethylenediaminetetraacetic acid (EDTA, BioUl- for cell lysis. The biocidal activity of the antibacterial poly- tra, anhydrous, ≥99%), and lysozyme (lyophilized powder, mer and the high surface area to volume ratio (SAVR) of the protein ≥90 %) were also purchased from Sigma-Aldrich but gold nanoparticles led to enhanced cell lysis. Unfortunately, were used without further purification. iTaq polymerase, 10X the modified gold nanoparticles also caused PCR inhibition; PCR buffer, and magnesium chloride were obtained from Bio- this was attributed to interactions between the nanoparticles Rad.(Montereal, QC Canada) Primers and dNTP mix (dATP, and DNA polymerase. dCTP, dGTP and dTTP) were obtained from Sigma-Aldrich. We recently described a new cell lysis method using an- Ethidium bromide (UltraPure 10 mg/mL, EtBr) and UltraPur tibacterial PPM columns fabricated within a 20% cross-linked Dithiothreitol (DTT, Cleland’s reagent) were purchased from poly(methyl methacrylate) (X-PMMA) microfluidic channel, Life Technologies Inc. (Burlington, ON Canada). A 1k bp that effectively lysed Escherichia coli (E. coli) and Bacil- DNA ladder was purchased from BioLabs (Ipswich, MA).

2 Table 1 Compositions of the eleven monolith solutions (in wt. %).a lith, before flushing with 250 µL of ethanol and 250 µL of deionized (DI) water to remove acid residues. 18 The depro- Monomers Cross-linkers tection process was validated by acquiring Fourier transform b Monolith BuMA Boc-AEMA EGDMA 1,6-HDDMA infrared spectroscopy (FT-IR) spectra for the porous column 1 1.3 15.6 9.1 - material. 18 2 1.3 15.6 8.1 1.0 3 1.3 15.6 7.1 2.0 4 1.3 15.6 6.1 3.0 Bacterial culture 5 1.3 15.6 5.1 4.0 E. coli DH5α and P. fluorescens (ATCC 13525) were served as 6 1.3 15.6 4.1 5.0 7 1.3 15.6 3.6 5.5 gram-negative test bacterial strains, while B. subtilis 168 and 8 1.3 15.6 3.1 6.0 E. saccharolyticus (ATCC 43076) were used as gram-positive 9 1.0 15.9 3.1 6.0 strains. E. coli DH5α and B. subtilis 168 were donated by Dr. 10 0.5 16.4 3.1 6.0 Michael Palmer (Chemistry Department, University of Water- 11 - 16.9 3.1 6.0 loo). E. saccharolyticus and P. fluorescens were purchased a All the reactions also included 21.5% cyclohexanol and 52.3% from American Type Culture Collection (ATCC, Manassas, 1-dodecanol as porogens, and 0.2% DMPAP as photoinitiator. VA, catalog numbers ATCC 43076 and ATCC 13525, respec- b aMonomer synthesis as described by Aly Saad Aly et al. 18. tively). Dgetails about bacterial culture and the growth conditions are provided in the Supporting Information.

Microchip fabrication Bacterial cell lysis The microfluidic channels were laser-micromachined within Samples of 2.5 mL each of the four bacterial cultures at con- 20% X-PMMA substrates with a 10.6 µm CO2 laser engrav- centration of 1.5 105, 2 105, 1.7 105 and 1.9 105 for for E. ing system (Universal Laser Systems, VLS2.30). The synthe- coli, B. subtilis, P. fluorescens and E. saccharolyticus, respec- et sis procedure for X-PMMA was described by Aly Saad Aly tively, were pelleted at 13k RPM for 3 min, washed twice with al 18 . The channels were 2.5 cm in length, 500 µm wide, and DI water, and then re-suspended in phosphate-buffered saline 250 µm deep. To obtain an enclosed channel another piece (PBS). of the X-PMMA substrate, with two holes drilled for the inlet and outlet, was chemically bonded with the substrate hosting On-chip cell lysis. A 110 µL aliquot of the bacterial sus- the microchannel by applying a thin layer of BuMA between pensions was pumped at a flow rate of 1.5 µL/min through the two X-PMMA layers, and placing the top and bottom sub- each of the monoliths prepared as described in Table 1, af- strates in a hot press under pressure (130 ◦C, 103 psi) for 30 ter deprotection, and the cell lysate was collected at the min (Heated Press 4386, Carver, Wabash, IN). Two 30G sy- biochip outlet to investigate the influence of the hydrophobic- ringe needles were trimmed and placed over the inlet and out- hydrophilic balance on the cell lysis efficiency of the antibac- let holes, and set with epoxy glue mixed with fine fumed silica terial monoliths. The bacterial suspension was also pumped powder to achieve a hard and stable adhesive. through monolith (11) before and after deprotection, at flow rates starting at 0.1 µL/min, and then increased in 0.4 µL/min Monolith formation and antibacterial activation increments after pumping 50 µL at each flow rate, to examine the influence of the flow rate on the cell lysis efficiency of the Mixtures with compositions as summarized in Table 1 were biochips. sonicated for 30 min to help dissolve the crystalline Boc- To invistigate the reusability of the biochips, a biochip was AEMA, stirred for 30 min under N2 flow, and then introduced reused 35 times and the cell lysis efficiency was evaluated for into the microchannel. Polymerization was triggered by irra- each cycle. The E. coli bacteria suspended in PBS used in this diation of the substrate for 15 min with a 365 nm UV source test were passed through the antibacterial monolith with the in a cabinet containing a UV lamp (200 mJ/cm2 intensity, optimal composition and at the optimal flow rate. Two differ- ENF-260C, Spectronics Corp. Westbury, NY). The substrate ent column washing procedures were compared for their in- was then turned over and irradiated with the UV source on fluence on the cell lysis efficiency. In the first procedure (PBS the other side for 15 min longer. Using a Pico plus syringe wash), the monolith was back-flushed after each run with 20 pump (Harvard apparatus, Holliston, MA), the microchannel µL of PBS buffer; after ten runs it was also washed with 25 L was then flushed with ethanol to remove the porogens and any of phosphoric acid and then with 20 µL of PBS. In the second unreacted monomer. Activation of the antibacterial monolithic procedure (Acid wash), the monolith was back-flushed after column was achieved through deprotection of the Boc-AEMA each use with 25 µL of phosphoric acid and then with 20 L units by flowing 250 µL of phosphoric acid through the mono- of PBS buffer. To investigate possible of DNA carryover, the

3 PBS recovered from the microfluidic channel in the back-flush the primers structure (Table S1), and the PCR cycles are pro- cycle was mixed with EtBr and any changes in fluorescence vided as Supporting Information. intensity were recorded. Gel electrophoresis. A Bio-Rad gel electrophoresis appa- Off-chip cell lysis. Aliquots of the bacterial suspensions ratus served to analyze the PCR products on 1.2% agarose gel, (300 µL) were pipetted into three centrifuge tubes. For ther- using a DC voltage of 85 V and a running time of 30 min. The mal cell lysis, one tube was immersed in a 90 ◦C water bath for gel was subsequently removed from the chamber and imaged three min and in a dry ice/acetone bath for three more min to with a Bio-Rad Doc XR imaging system. create a thermal shock. This freeze-thaw cycle was repeated three times. More details about the off-chip mechanical and Results and discussion chemical cell lysis protocols are provided in the Supporting Information. Influence of the hydrophobic-hydrophilic balance It was previously determined that the hydrophobic- Cell Lysis efficiency hydrophilic balance is a critical parameter controlling the activity of antibacterial polymers (SMAMPs). 19 Thus we The bacterial cell lysates collected at the outlet of the biochips investigated the effect of varying the amphipathic nature of described in Table 1 were analyzed to study the influence of the antibacterial monoliths on their cell lysis efficiency, by the hydrophobic-hydrophilic balance and the flow rate on the changing the proportions of the hydrophobic (BuMA) and cell lysis efficiency of the monoliths by the methods described the amine-containing hydrophilic (Boc-AEMA) monomers, below. as well as the cross-linking hydrophobic (1,6-HDDMA) and hydrophilic (EGDMA) monomers, as outlined in Table 1). DNA detection by fluorometry. The ethidium bromide The ethidium bromide (EtBr) intercalation assay sevred to (EtBr) intercalation assay was used as an indicator of the pres- detect DNA in the cell lysate, as evidence for cell lysis. When ence of DNA in the cell lysate. A 0.02 mL aliquot of the ethidium bromide (EtBr) is exposed to UV light at 285 nm it bacterial cell lysate was added to a spectrofluorometer cuvette fluoresces with an orange color at 595 nm, which intensifies containing 0.380 mL of DI water and 0.03 mL of EtBr from a considerably after its intercalation in DNA. The fluorescence stock solution with a concentration of 0.4 mg/L, and the flu- intensity for EtBr before and after intercalation in the DNA orescence intensity of EtBr was measured at a wavelength of present in the cell lysate collected at the outlet of the porous 595 nm measured on a Quanta-Master 4 spectrofluorometer antibacterial monoliths (1-11) was quantified on a spectroflu- (Photon Technology International, London, ON Canada). orometer to obtain Fig. 1. DNA concentration by UV-Vis spectrophotometry. A 4 As can be seen from Fig. 1, the fluorescence intensity for µL aliquot of PBS buffer was pipetted onto the end of a fiber the tested monoliths can be divided into three groups: Groups optic cable (receiving fiber) of a UV-Vis Spectrophotome- I (columns 1 through 5), II (columns 6 through 8), and III ter NanoDrop 2000c (Thermo Scientific, Mississauga, ON (columns 9 through 11). In Group I, the fluorescence in- Canada)to serve as blank (reference). A second fiber optic tensity of EtBr gradually increases as the content in the hy- cable (source fiber) was then brought into contact with the liq- drophobic (1,6-HDDMA) and hydrophilic (EGDMA) cross- uid sample, causing the liquid to bridge the gap between the linking monomers increase and decrease within that series, fiber optic ends. Then a 4 µL sample of the E. coli DH5α, B. respectively. This reflects increasing DNA concentrations in subtilis 168, P. fluorescens (ATCC 13525), or E. saccharolyti- the crude lysates collected at the outlet of the microfluidic cus (ATCC 43076) bacterial cell lysate was separately pipetted channels hosting the monoliths, and therefore enhanced lysis onto the receiving fiber, to measure the DNA concentration in as the columns become increasingly hydrophobic. In Group the lysate after flowing the cell suspension through both the II, the fluorescence intensity of EtBr is relatively insensitive protected and deprotected columns. to further variations in hydrophobic and hydrophilic cross- linking monomers (columns 6-8). This saturation shows that PCR reagents and experimental setup. The PCR reaction further increasing the hydrophobicity of the monoliths does was performed in a T100 Thermal Cycler (Bio-Rad, Montreal, not lead to improved lysis ability for the bacterial species QC Canada) in a 25 µL volume consisting of 300 nM of for- tested in this study. In Group III, the fluorescence intensity ward primer, 300 nM of reverse primer, 200 µM of dNTPs, of EtBr again increases as the contents of the hydrophobic 3.5 mM of magnesium chloride, 0.625 U of iTaq polymerase, (BuMA) and hydrophilic, positively charged (AEMA) non- 2.5 µL of 10X PCR buffer, and 200 ng of DNA present in the cross-linking monomers decrease and the increase, respec- crude cell lysate collected at the outlet of the biochip. More tively, in the last monoliths. This increase could be ibuted to details about the bacterial cultures and the growth conditions, further changes in the hydrophobic-hydrophilic balance, but

4 Fig. 1 Florescence intensity of EtBr after intercalating into the DNA released from bacterial cells ﬂowing through the microﬂuidic channels hosting the different PPM columns. Fig. 2 Absolute concentration of DNA present in the crude cell lysate collected at the outlet of the different monoliths.

(on the basis of the trends described above) it is more likely due to increased charge density in the monolith as a result pension was pumped at a flow rate of at least 5 µL/min for the of the higher concentration of the positively charged amine specific system they used. In the current study, we investigated monomer (AEMA). the influence of the flow rate on both mechanical shearing and contact killing for bacterial cell lysis. Control samples were To validate the cell lysis results obtained by the EtBr as- prepared for the ethidium bromide intercalation assay contain- say and to directly quantify the lysis efficiency of the differ- ing 0.02 mL of bacterial cell suspensions that were not flown ent columns, the DNA concentration in the cell lysates col- through the monolith, 0.380 mL of DI water, and 0.03 mL of lected at the exit of the microfluidic channels was determined EtBr from a stock solution with a concentration of 0.4 mg/L. by UV-Vis spectrophotometry as shown in Fig. 2. It can be seen that the DNA concentration in the different cell lysates Protected PPM Column. Before removing the Boc pro- also matches the three regions identified in Fig. 1. This further tecting group from the Boc-AEMA monomer, the concentra- confirms that column (11) had the highest antibacterial activ- tion of DNA in the cell lysates collected from column (11) ity among the different monolith columns investigated; thus it varied with the flow rate according to four different regimes as was used for the subsequent experiments. Fig. 1 and Fig. 2 shown in Fig. 3. At flow rates below 1.2 µL/min (Regime i) both show that the biochips display a comparable lysis effi- the fluorescence intensity for EtBr was similar to the control ciency for P. fluorescens and E. coli ( the two gram-negative sample, indicating that no significant lysis took place. This bacteria), as well as for E. saccharolyticus and B. subtilis (the implies that the cells cannot be mechanically lysed through gram-positive bacteria). However it is clear that the lysis effi- the porous medium of the monolith unless the flow rate ex- ciency is significantly higher for the gram-negative than for ceeds a certain threshold value, in agreement with the findings the gram-positive bacteria within the monolith composition of Burke et al. At flow rates between 1.2 and 4.8 µL/min range investigated. This is reasonable since gram-positive (Regime ii) the fluorescence intensity of EtBr (and the DNA bacteria have a thicker cell membrane than gram-negative bac- concentration) increased rapidly, as shown in Fig. 3 (and Fig. teria, which makes them harder to lyse. S1, in the Supporting Information). However at flow rates between 5.2 and 7.6 µL/min (Regime iii) the fluorescence inten- Influence of flow rate sity of EtBr and the DNA concentration gradually leveled off (decreasing slope). In these two regimes, as the flow rate in- The bacterial cell wall/membrane is mechanically sheared by creases, the bacterial cells are forced faster through the porous flowing through the porous medium of the monolith, but it medium of the monolith, which results in more mechanical is also damaged and disintegrated by physical contact with lysis of the cells. Finally, at flow rates between 6.7 and 10 the antibacterial polymeric biocide covering the porous sur- µL/min (Regime iv) the fluorescence intensity of EtBr and face. Both effects lead to leakage of the intracellular content. the DNA concentration saturated which shows that there was Burke et al. 13 thus reported that the flow rate was a critical no influence of the flow rate on cell lysis efficiency beyond factor in the mechanical shearing of B lymphocyte cells in 6.7 µL/min. The influence of the flow rate on cell lysis was porous monolith columns. They demonstrated that B lympho- therefore highest in Regime ii, less significant in Regime iii, cyte cells could only be mechanically lysed when the cell sus- and insignificant in Regime iv. The results also reveal that

5 Fig. 3 Florescence intensity of EtBr after intercalating into the DNA released from bacterial cells pumped through the protected PPM column (11) at different flow rates. the lysis efficiency was higher for gram-negative than gram- is still taking place (4-7.6 µL/min) however, until the flow positive bacteria within the flow rate range investigated. This rate reaches a value (8 µL/min) where further increasing the was again expected, as gram-positive bacteria are harder to flow rate has no effect on both lysis mechanisms. The highest lyse due to their thicker cell membranes. overall cell lysis efficiency (highest overall DNA concentration and fluorescence intensity) is therefore achieved at flow Deprotected (antibacterial) PPM column. After depro- rates between 4 and 6 µL/min, as a result of a combination of tecting the amine group of the Boc-AEMA monomer, the con- mechanical shearing and contact-killing mechanisms. centration of DNA in the cell lysates was found to vary with the flow rate in five different regimes, as shown in Fig. 4. At Reusability of the biochip flow rates below 1.2 µL/min (Regime i) the DNA concentration are similar to the control samples, indicating that no lysis The change in EtBr fluorescence intensity due to its interca- happened. This further confirms the results gathered for the lation in the DNA present in the cell lysate is compared in protected monolith. Fig. 5 after each injection and flushing cycle for the two dif- At flow rates between 1.2 and 3.6 µL/min (Regime ii), ferent washing protocols examined. It can be seen that for the the DNA concentration increased rapidly.The DNA concen- PBS washing protocol the antibacterial efficiency decreases tration (and The fluorescence intensity of EtBr) at a flow slightly over 10 cycles, but washing of the monolith with phos- rate of 1.2 µL/min, shown in Fig. 4 and Fig. S2 (Support- phoric acid then restores the activity to some extent. This ing Information), are 13 times higher and more than twice could be due to gradual deprotonation of the amine caused by higher, respectively, than the values obtained before deprotec- the sequential PBS washes, which is being protonated again by tion (Fig. 3 and S1). This confirms that additional lysis took the phosphoric acid wash. The decrease in activity observed place after deprotection due to the antibacterial nature of the for the Acid wash protocol is much more gradual. It is inter- monolithic surface. At flow rates between 4 and 6 µL/min esting to note that after the twentieth use, the lysis efficiency (Regime iii), the DNA concentration (and fluorescence in- started degrading dramatically regardless of the washing pro- tensity of EtBr)saturated. This contrasts with the trend ob- tocol used. The monolith was completely blocked after 30-35 served before deprotection within the same flow rate range, cycles. Gradual blockage of the pores by cell debris, leading to since the fluorescence intensity of EtBr and the DNA concen- a decrease in the surface area of the porous monolith column tration continued to increase. At flow rates between 6.4 and accessible to the cells, may therefore also explain in part the 7.6 µL/min (Regime iv) the fluorescence intensity of EtBr and gradual decrease in cell lysis efficiency observed over multiple the DNA concentration both decreased rapidly, while the flu- cycles. Irrespective of the washing protocol used, the perfor- orescence intensity of EtBr and DNA concentration continued mance of the biochip appears acceptable over multiple cycles to increase over the same flow rate range for the columns in since the lysis efficiency only decreased by 10% over 20 cy- the protected state. At flow rates between 8 and 10 µL/min cles. The florescence intensity of EtBr mixed with the PBS (Regime v) the fluorescence intensity of EtBr and the DNA recovered from the microfluidic channel in the back-flush cy- concentration levelled off again. cle showed insignificant DNA carry over, reaching only 0.6% These results show that when the flow rate exceeds a criti- of the maximum EtBr intensity reported as shown in Table cal value the bacterial cells spend insufficient time in contact S2 (Supporting Information). It is also worth mentioning that with the antibacterial surface of the monolith, which leads to the total cycle time for the biochip was 35 min including both less efficient contact-killing lysis. Mechanical shearing lysis sample lysis and monolith regeneration.

6 Fig. 4 Concentration of the DNA present in the crude bacterial cell lysate collected at the outlet of the microfluidic channels hosting deprotected PPM column (11) flown at different flow rates.

Fig. 5 Fluorescence intensity for EtBr intercalated into the DNA released from bacterial cells ﬂowing through the protected monolith (11) at 4 µL/min over successive runs for the two different washing Fig. 6 Concentration of DNA in the crude bacterial cell lysate protocols (PBS and Acid). collected at the outlet of monolith (11) at a ﬂow rate of 4 µL/min, and after mechanical, thermal and chemical cell lysis.

Biochip vs. off-chip cell lysis

A 100 µL aliquot of the bacterial suspension (at the same concentration used in the monolith lysis experiments) was lysed by the off-chip cell lysis methods. The efficiency of the different cell lysis techniques, including the biochip method, is compared in Fig. 6. It is clear that the DNA concentration trophoresis. The analysis results are shown in Fig. 7 for the in the cell lysate collected from the biochip was higher than PCR products of E. saccharolyticus (column 2), B. subtilis for the off-chip (mechanical, thermal and chemical) methods. (column 5), P. fluorescens (column 7), and E. coli (column 9) This shows that the biochip approach led to more efficient ly- bacterial cells that were not passed through the porous mono- sis of the bacterial species tested than the traditional off-chip lith column, and the PCR products of E. saccharolyticus (col- techniques. Furthermore, the biochip lysis method does not umn 3), B. subtilis (column 4), P. fluorescens (column 6), and require power consumption (apart from the pump operation), E. coli (column 8) bacterial cell lysates collected at the outlet chemical/enzymatic reagents, the use of centrifugation, soni- of the deprotected monolith. There are no detectable amounts cation, nor a complicated design and fabrication process. As of DNA at the PCR output for the bacteria samples that were a result, the on-chip cell lysis technique developed appears not lysed in the monolithic column, which confirms that the well-suited for incorporation in an integrated sample prepara- cells had intact membranes before passing through the mono- tion system. lith. In contrast, DNA is clearly detected at the PCR output for the samples that were passed through the deprotected mono- PCR and gel electrophoresis lith, which shows that the membrane of cells was disintegrated by flowing through the monolith. This is a clear sign for lysis, The genes in the DNA released by lysing the bacterial cells but also demonstrates that the antibacterial monoliths did not were amplified by PCR and qualitatively validated by gel elec- leach any PCR inhibitors.

7 Acknowledgements

We would like to express our gratitude to Grand Challenges Canada and Bigtec Labs for their ﬁnancial support. We also like to thank Prof. Michael Palmer, Mr. Mohamed Salah, Mr. Eric K. Brefo-Mensah, and Mr. Olivier Nguon from the Chemistry Department at the University of Waterloo, for pro- viding E. coli and B. subtilus bacterial samples, access to their lab facilities and their guidance on some of the biological and chemical concepts.

Supporting information Additional ﬁgures and information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Fig. 7 Gel electrophoresis analysis of PCR output for unlysed References (column 2) and lysed (through monolith) (column 3) E. 1 D. W. Lee and Y.-H. Cho, Sens. Actuat. B-Chem., 2007, 124, 84 – 89. saccharolyticus; unlysed (column 5) and lysed (through monolith) 2 D. C. Chang, B. M. Chassy, J. A. Saunders and A. E. Sowers, Guide to (column 4) B. subtilis; unlysed (column 7) and lysed (through Electroporation and Electrofusion, Academic Press, San Diego, 1992. monolith) (column 6) P. fluorescens; and unlysed (column 9) and 3 J. Sambrook and D. W. Russel, Molecular Cloning: A Laboratory Man- lysed (through monolith) (column 8) E. coli. Column 1 is for a 1k ual, Cold Spring Harbor, New York, 2001. bp DNA ladder. 4 C.-Y. Lee, G.-B. Lee, J.-L. Lin, F.-C. Huang and C.-S. Liap, J. Mi- cromech. Microeng., 2004, 15, 1215–12233. 5 D. Irimia, R. G. Tompkins and M. Toner, Anal. Chem., 2004, 76, 6137 – 6143. 6 M. Cichova, M. Proksova, L. Tothova, H. Santha and V. Mayer, Cent. Eur. Conclusions J. Biol., 2012, 7, 230–240. 7 P. Sethu, M. Anahtar, L. L. r. Moldawe, R. G. Tompkins and M. Tone, Anal. Chem., 2001, 76, 6247 – 6253. 8 J. El-Ali, S. Gaudet, A. Gunther, P. K. Sorger and K. F. Jensen, Anal. Microfluidic biochips were fabricated that have the ability Chem., 2005, 77, 3629 – 3636. 9 E. A. Schilling, A. E. Kamholz and P. Yager, Anal. Chem., 2002, 74, 1798 to efficiently lyse four species of gram-positive and gram- – 1804. negative bacteria: E. saccharolyticus (ATCC 43076), B. sub- 10 Y.-B. Kim, J.-H. Park, W.-J. Chang, Y.-M. Koo, E.-K. Kim and J.-H. Kim, tilis 168, E. coli DH5, and P. fluorescens (ATCC 13525). The Biotechnol. Bioproc. E., 2006, 11, 288 – 292. lysis ability of the biochips was validated with the ethidium 11 J. Hall, E. Felnagle, M. Fries, S. Spearing, L. Monaco and A. Steele, bromide intercalation assay, relating the presence of DNA in Planet. Space Sci., 2006, 54, 1600 – 1611. 12 D. D. Carlo, K.-H. Jeong and L. P. Lee, Lab Chip, 2003, 3, 287–291. the cell lysate with an increase in fluorescence intensity for 13 J. M. Burke and E. Smela, Biomicrofluidics, 2012, 6, 016506–1–016506– EtBr, and UV-Vis spectrophotometry to directly determine the 10. DNA concentration in the cell lysate. Gel electrophoresis 14 M. Mahalanabis, H. Al-Muayad, M. D. Kulinski, D. Altman and C. M. analysis of the PCR products after DNA gene amplification Klapperich, Lab Chip, 2009, 9, 2811–2817. of the cell lysate showed that the monoliths did not leach any 15 K. Lienkamp, A. E. Madkour, A. Musante, C. F. Nelson, K. Nusslein and G. N. Tew, J. Am. Chem. Soc., 2008, 130, 9836–9843. material interfering with the PCR process. The performance 16 J. C. Tiller, C.-J. Liao, K. Lewis and A. M. Klibanov, J. Am. Chem. Soc., of the microfluidic biochips developed exceeded that of the 2001, 98, 5981–5985. traditional off-chip mechanical, thermal, and chemical cell ly- 17 W. Wan and J. T. Yeow, Biomed. Microdevices, 2012, 14, 337–346. sis techniques. The influence of the hydrophobic-hydrophilic 18 M. Aly Saad Aly, O. Nguon, M. Gauthier and J. T. Yeow, RSC Adv., 2013, balance on the lysis efficiency was investigated, and the an- 3, 24177–24184. tibacterial monolith with the highest lysis efficiency was used 19 A. Rao, Arch. Biochem. Biophys., 1999, 361, 127 – 134. to determine the influence of the flow rate of the bacterial suspension through the porous monolith. It was also shown that Graphical TOC Entry the biochips can be reused for at least twenty times without significant performance degradation or carryover when they are back-flushed between cycles.

8 9